Updated: Dec 10, 2024
base modelDownload
1 variant available
fp16 SafeTensor
VixenMix-SD1.5_8StepsHYPER_LoRA_Merge.safetensors
Half precision, best balance (pruned) • 2.14 GB
Verified: 2 years ago
1,2900 1 2 3 4 5 6 7 8 9,0 1 2 3 4 5 6 7 8 90 1 2 3 4 5 6 7 8 90 1 2 3 4 5 6 7 8 9
(39)
Sep 26, 2024

21.6K0 1 2 3 4 5 6 7 8 90 1 2 3 4 5 6 7 8 9.0 1 2 3 4 5 6 7 8 9K
109.6K0 1 2 3 4 5 6 7 8 90 1 2 3 4 5 6 7 8 90 1 2 3 4 5 6 7 8 9.0 1 2 3 4 5 6 7 8 9K
632.2K0 1 2 3 4 5 6 7 8 90 1 2 3 4 5 6 7 8 90 1 2 3 4 5 6 7 8 9.0 1 2 3 4 5 6 7 8 9K
License:
CreativeML Open RAIL++-MFirst and foremost - this model was uploaded without a Hyper merge by: jjambong000
Link to the original model: VixenMix v1.0
I have just merged the Hyper SD 1.5 (8 Steps) LoRA into it, with a specific ratio. The results were too nice to not upload the model, especially at just 8 steps (excluding the hires/upscale). This was mainly made for a bounty submission, requested by marcus199911.
VAE is baked in.
Note the following gen parameters and use them for the best results:
Sampler: DPM++ SDE (recommended)
Scheduler: sgm_uniform (IMPORTANT)
Steps: 8 (recommended)
CFG: 1.5 - 2
Hires fix or UltimateSDUpscale: Denoise at 0.28, 4x-UltraSharp, dpm++ sde, sgm_uniform at 1.7x.
If you want to support more models like this, or you just like what you see: ko-fi (。♥‿♥。)
Also, I don't plan on charging buzz for my LoRAs. So, I'd really appreciate if by any chance you could tip the model.

